* Table 3: Determinants of survey attitudes

eststo: quietly: svy: reg enjoyfactor i.female c.respage respeduc ib1.region5
eststo: quietly: svy: reg valuefactor i.female c.respage respeduc ib1.region5
eststo: quietly: svy: reg reliabilityfactor i.female c.respage respeduc ib1.region5
eststo: quietly: svy: reg costsfactor i.female c.respage respeduc ib1.region5
eststo: quietly: svy: reg privacyfactor i.female c.respage respeduc ib1.region5
eststo: quietly: svy: reg intentionsfactor i.female c.respage respeduc ib1.region5
esttab using Survey.Attitudes.Table.3.rtf, title(Determinants of Survey Attitudes, by Geographical-Cultural Region) replace p r2 star(+ 0.10 * 0.05 ** 0.01 *** 0.001) 
eststo clear

* Table 4: Cultural-geographical group differences in survey attitudes

svy: reg enjoyfactor ib1.region5
margins, by(region5)

svy: reg valuefactor ib1.region5
margins, by(region5)

svy: reg reliabilityfactor ib1.region5
margins, by(region5)

svy: reg costsfactor ib1.region5
margins, by(region5)

svy: reg privacyfactor ib1.region5
margins, by(region5)

svy: reg intentionsfactor ib1.region5
margins, by(region5)

* Table 6: Marginal effects of attitudes on continuation

svy: logit bday c.reliabilityfactor##i.region3

margins, at(reliabilityfactor=(-1 1)) by(region3) post coefl
test  _b[1bn._at#0bn.region3] = _b[2._at#0bn.region3]
* And so on for others

* Table A8: Bivariate logit coefficients

svy: logit bday c.enjoyfactor if region3==0
svy: logit bday c.enjoyfactor if region3==1
svy: logit bday c.enjoyfactor if region3==2
* And so on for other dimensions

* Figure 4: Effect of survey reliability on continuation, by group

svy: logit bday c.reliabilityfactor##i.region3
margins, at(reliabilityfactor=(-3(1)1)) by(region3) 
marginsplot, xtitle("Perceived Survey Reliability") scheme(s1mono) ytitle("Probability of Continuation (Birthday)") legend(cols(3)) plotopts(msize(small)) title("")





